A Multi-Sensor Based Missile Target Type Recognition Method

Abstract

The invention discloses a guided missile target type recognizing method based on multiple sensors. The method comprises following steps: firstly, the sensors are used for collecting guided missile target data, and according to the guided missile target data collected by the sensors, a training sample matrix and a test sample matrix are built; the training sample matrix is subjected to singular value decomposition, and according to the contribution rate of the singular value, dimension reduction dimensions are determined; thirdly, according to a sample matrix obtained through singular value decomposition, a classification model is generated, and the classification model is composed of a linear transformation matrix and a classifying center vector set; and fourthly, a sample to be classifiedis subjected to projection pretreatment, and sample vectors to be classified are obtained, the cosine distance between the vectors and each component in the classification center vector set obtainedin the third step is calculated, and the set with the shortest distance serves as the type recognizing result. The multiple sensors can be used for generating high-dimension sample data, and guided missile target type recognizing considering efficiency and accuracy can be carried out.

Description

technical field [0001] The invention relates to ship combat system technology, in particular to a multi-sensor-based missile target type identification method. Background technique [0002] In ship-to-air self-defense operations, different types of incoming missiles have different flight speeds, angles of incidence, and terminal maneuvering modes. Therefore, different defensive operations must be adopted for different types of missiles. Identifying incoming missile target types can improve the pertinence of defensive operations. [0003] The requirements for the recognition accuracy of missile target types in ship-to-air self-defense systems are increasing. The traditional method of using a single source of information to aid identification can no longer meet the needs of combat use, and the use of multi-sensors for multi-source information fusion to achieve accurate identification of missile target types has become a current research hotspot. However, the data generated b...

AI Technical Summary

Problems solved by technology

The traditional method of using a single source of information to assist identification can no longer meet the needs of combat use, and the use of multi-sensors for multi-source information fusion to achieve accurate identification of missile target types has become a current research hotspot

However, the data generated by multi-sensors has a high dimensionality, which not only contains the characteristic information of missile targets, but also contains redundant and irrelevant information.

In a high-dimensional space, the distinguishability of distances between different samples decreases as the dimensionality of sample data increases. High-dimensional sample data not only brings exponential growth in the complexity of classification calculations, but also redundancies between dimensions and irrelevant information also reduces the accuracy of the classification algorithm

Therefore, the effect of directly using the data generated by multiple sensors to identify missile target types is not good

Images